Scholarlink

<p>The composition of municipal solid waste incineration fly ash is influenced by regional waste characteristics, incineration technology, and air pollution control systems. However, systematic characterization of the elemental distribution patterns across large-scale databases remains limited. Therefore, this study constructed a comprehensive dataset of 1439 fly ash samples from multiple countries and Chinese provinces and used integrated statistical analysis, machine learning imputation, and correlation network approaches to reveal the underlying geochemical relationships. Random forest imputation successfully addressed missing data problems (11.7%–65.3% missing rates) while preserving the correlation structures. K-means clustering only produced 2–3 clusters with weak differentiation (silhouette: 0.345–0.378; adjusted rand index: 0.024–0.034). Gaussian Mixture Model selection showed no strongly preferred discrete component numbers; and t-distributed stochastic neighbor embedding visualization confirmed extensive geographic overlap across all perplexity settings (5–100). These results demonstrated that elemental composition followed a continuous spectrum. An element correlation network analysis revealed small-world architecture (density: 0.427, clustering: 0.669), with the hub rankings partially sensitive to geographic sampling composition under balanced subsampling. There were three thermochemically coherent element groups, refractory matrix-forming elements (Al-Si-Fe-Ti), volatile heavy metals (Pb-Zn-Cu-Cd), and soluble salts (Ca-K-Na-Cl), linked to distinct physicochemical mechanisms. The regional networks analysis showed distinct connectivity patterns (density: 0.380–0.602) with region-specific hub identities; community detection identified major element-heavy metal segregation (modularity: 0.175–0.368); and the robustness analysis demonstrated that distributed multi-hub structures improved resilience. Differentiated treatment strategies were proposed based on regional network vulnerability profiles, providing mechanistic guidance for region-specific fly ash resource recovery.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Network analyses reveal geographically continuous elemental patterns in municipal solid waste incineration fly ash

  • 期刊论文

  • 发表日期：

    2026年7月16日

查看全文
  • Khin Thant Sin,
  • Siyuan Yu,
  • Ke Chang,
  • Pinjing He,
  • Fan Lü,
  • Hua Zhang

摘要

The composition of municipal solid waste incineration fly ash is influenced by regional waste characteristics, incineration technology, and air pollution control systems. However, systematic characterization of the elemental distribution patterns across large-scale databases remains limited. Therefore, this study constructed a comprehensive dataset of 1439 fly ash samples from multiple countries and Chinese provinces and used integrated statistical analysis, machine learning imputation, and correlation network approaches to reveal the underlying geochemical relationships. Random forest imputation successfully addressed missing data problems (11.7%–65.3% missing rates) while preserving the correlation structures. K-means clustering only produced 2–3 clusters with weak differentiation (silhouette: 0.345–0.378; adjusted rand index: 0.024–0.034). Gaussian Mixture Model selection showed no strongly preferred discrete component numbers; and t-distributed stochastic neighbor embedding visualization confirmed extensive geographic overlap across all perplexity settings (5–100). These results demonstrated that elemental composition followed a continuous spectrum. An element correlation network analysis revealed small-world architecture (density: 0.427, clustering: 0.669), with the hub rankings partially sensitive to geographic sampling composition under balanced subsampling. There were three thermochemically coherent element groups, refractory matrix-forming elements (Al-Si-Fe-Ti), volatile heavy metals (Pb-Zn-Cu-Cd), and soluble salts (Ca-K-Na-Cl), linked to distinct physicochemical mechanisms. The regional networks analysis showed distinct connectivity patterns (density: 0.380–0.602) with region-specific hub identities; community detection identified major element-heavy metal segregation (modularity: 0.175–0.368); and the robustness analysis demonstrated that distributed multi-hub structures improved resilience. Differentiated treatment strategies were proposed based on regional network vulnerability profiles, providing mechanistic guidance for region-specific fly ash resource recovery.